Combining phytoliths and δ13C matter in Holocene palaeoenvironmental studies of tropical soils: an example of an Oxisol in Brazil

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/copyright

Author's personal copy

Combining phytoliths and d13C matter in Holocene palaeoenvironmental studiesof tropical soils: An example of an Oxisol in Brazil

Marcia R. Calegari a,*, Marco Madella b, Pablo Vidal-Torrado c, Luiz Carlos R. Pessenda d,Flávio A. Marques e

aDepartamento de Geografia, Universidade Estadual do Oeste do Paraná, MCR, Rua Pernambuco, 1777, Caixa Postal 91, CEP 85970-020, Mal. Cd. Rondon, Paraná, Brazilb ICREA, Department of Archaeology and Anthropology, Institució Milà i Fontanals, Spanish National Research Council (CSIC), C/Egipcíaques, 15, 08001 Barcelona, SpaincDepartamento de Ciência do Solo, Escola Superior de Agricultura Luiz de Queiroz (ESALQ, Universidade de São Paulo), Av. Pádua Dias, 11, Caixa Postal 9, CEP 13418-900,Piracicaba, São Paulo, Brazild Laboratório de 14C, Centro de Energia Nuclear na Agricultura (CENA), Av. Centenário 303, Caixa Postal 96, CEP 13400-000, Piracicaba, São Paulo, Brazile Embrapa Solos /UEP, Rua Antônio Falcão 402, Boa Viagem CEP, 51020-240, Recife, Pernambuco, Brazil

a r t i c l e i n f o

Article history:Available online 18 November 2011

a b s t r a c t

Many plants deposit the soluble silica absorbed from the soil as monosilicic acid (H4SiO4) in and betweentheir cells, generating bodies of opal silica (SiO2$nH2O) called phytoliths. Although phytoliths aresusceptible to dissolution under extreme pH conditions, they generally do remain in the soil for longperiods of time and can help in the reconstruction of past vegetation and climates. In the present study,phytolith analysis was used to reconstruct the palaeoenvironmental conditions that contributed to thepedogenetic processes, the deposition of organic matter and its stabilization in a very thick (>1 m)umbric epipedon of a Humic Hapludox profile from Minas Gerais State (Brazil). The results from thephytolith assemblages were also compared to the fractions and isotopic data of soil carbon of the sameprofile. The result from studying these two palaeoenvironmental proxies together has shown that theenvironment under which the umbric epipedon was formed was a mixture of vegetation withpredominance of C3 plants in mesothermic conditions and with little variation in humidity since MiddleHolocene.

� 2011 Elsevier Ltd and INQUA.

1. Introduction

In tropical and sub-tropical areas ofBrazil, there areHumicOxisolswith thick epipedons (A horizon >100 cm) overlying the diagnosticBwhorizons. The epipedons have dark colour (value and chroma< 4)and significant content of organic carbon in the deeper layers. Thegenesis and palaeoenvironmental significance of these soils are notcompletely understood. Most authors have suggested that they arerelic soils in landscapes that had a favourable climate for organiccarbon accumulation (Nakashima, 1973; Queiroz Neto and Castro,1974; Kampf and Klamt, 1978; Lepsch and Buol, 1986; Silva andVidal-Torrado, 1999). Carbon was then preserved due to severalspecific soil and environmental factors such as high acidity, low-basesaturation, relatively cold climate, stable geomorphic surfaces andaccumulation of charred material (charcoal) that had been partiallyaltered, decomposed into microparticles, and distributed in the soilby biological activity. The study of Humic Oxisols with thick

epipedons is particularly interesting because of the significantamountof accumulated organic carbon (300 t/ha in thefirst 1m),fivetimes higher than in other Oxisol classes, providing the potential tostudypalaeoenvironmental reconstruction inBrazilian tropical areas.In many of these Oxisols, some considered as polycyclic soils (Lepschand Buol, 1986), past climate changes may have been recorded.

The intensity of climate variability occurring in Brazil betweenglacial and interglacial periods influenced the weathering andpedogenesis rates of soils as well as the regional floristic coverdistribution. This took place throughout the Quaternary by quickand recurring expansions and retractions of forests at the expenseof more open vegetation such as the Campo (grassland savannah),the Caatinga (open arboreal savannah) and the Cerrado (closedarboreal savannah) (Martinelli et al., 1996; Pessenda et al., 1996,1998, 2005). There have been several isotopic and pollen studiesaimed at defining the boundaries between forest and open vege-tation (see for example, Absy et al., 1991; Oliveira, 1992; Ledru et al.,1994; Pessenda et al., 1996, 2004; Gouveia et al., 1997). Pollenstudies, d13C, and 14C dating provided evidence that in northeastand southeast Brazil, climate oscillations during the Quaternarywere not always synchronous (Scheel-Ybert et al., 2003).

* Corresponding author.E-mail address: [email protected] (M.R. Calegari).

Contents lists available at SciVerse ScienceDirect

Quaternary International

journal homepage: www.elsevier .com/locate/quaint

1040-6182/$ e see front matter � 2011 Elsevier Ltd and INQUA.doi:10.1016/j.quaint.2011.11.012

Quaternary International 287 (2013) 47e55

Author's personal copy

The Humic Oxisols (Latossolos húmicos) cover an area of144,000 km2 of the Brazilian territory (FAO-UNESCO, 1981). Theyare most commonly found in the south and southeast (tropical andsub-tropical environments) on elevated topography, indicative ofancient erosional surfaces, and show an abundant presence ofmacroscopic soil charcoals at great depth in the profile, togetherwith high contents of organic carbon in the umbric epipedon. Thesecharacteristics are not compatible with the current warmer andmore humid climate (Lepsch and Buol, 1986). In this respect, thehyper-developed umbric epipedon can be seen as an importantcarbon reservoir and a characteristic feature of the soil. The study ofthis soil type is of fundamental importance to understand thepedogenetic processes in tropical and sub-tropical areas of theSouth American continent as well as changes in vegetation cover.

Opal phytoliths are a potential source of information forreconstructing structural changes in vegetation and the edaphicconditions of soil formation through time, and can be important inoxic environments, offsetting the paucity of fossil pollen (Argant,1990; Alexandre et al., 1997; Bremond et al., 2005). Soils in trop-ical areas generally show much higher silica concentrations(of which phytoliths are an important proportion) than soils intemperate areas (Piperno, 2006). In Brazil, however, there aresurprisingly very few phytolith studies. Kondo and Iwasa (1981)investigated the phytolith assemblages of an Oxisol, locallycalled Terra Preta de Índio (black earth-like anthropogenic soil) in

the Amazon region, detecting changes in the environmentalconditions during the formation of the soils. Alexandre et al.(1999) studied the phytolith assemblage of an Oxisol situated atthe boundary between forest and Cerrado in Salitre (Minas GeraisState) in southeastern Brazil. The phytolith analysis, together withisotope data from soil organic matter, highlighted a savannahphase (drier climate during the middle Holocene between 5500and 4500 BP) that was followed by two periods with moredeveloped tree communities, the first between ca. 4000 and 3000BP and the second after ca. 970 BP. Borba-Roschel et al. (2006)published the results of a comparison between phytolith assem-blages obtained from the modern vegetation and a phytolithsequence obtained from a peat profile located in a Brazilian Cer-rado at Uberaba (Minas Gerais State). Given the potential that opalphytoliths and d13C offer to palaeoenvironmental reconstruction,the objective of this study was to analyze the combined records toreconstruct the palaeovegetation, and palaeo-edaphic/climaticcondition that lead to the development of a very thick umbricepipedon in Brazilian oxisols.

2. Regional setting

For this study, a representative Humic Oxisol pedon class, verydeep and rich in organic carbon, classified as Humic Hapludoxwas selected. The soil (Fig. 1) is located in Machado, southern

Fig. 1. Humic Hapludox profile showing the very thick umbric epipedon, (Machado, Minas Gerais State, Brazil).

Table 1Sample list, soil properties of each examined sample.

Horiz. Depth cm Munsell colour Clay Silt Sand C-Total Bulk density pH (H2O)Fine Coarse

g kg�1 t m�3 (1:2.5)

Ap 0e10 5 YR 3/2 455 139 142 264 47.5 0.8 4.2A2 10e20 53.7

20e30 83.830e40 5 YR 2.5/2 498 176 173 153 57.3 0.8 4.740e50 39.450e60 40.0

A3 60e70 32.970e80 25.580e90 5 YR 2.5/2 567 131 140 162 28.4 0.9 5.0

AB 90e100 26.5100e110 5 YR 2.5/2 540 176 120 164 23.5 1.0 5.0110e120 0 22.9

BA 120e140 5 YR 4/4 598 147 137 119 18.7 1.1 4.9Bw1 140e170 5 YR 4/6 624 136 134 107 12.5 1.1 5.5Bw2 170e200þ 5 YR 5/8 650 88 151 112 9.3 1.0 5.3

M.R. Calegari et al. / Quaternary International 287 (2013) 47e5548

Author's personal copy

Minas Gerais State, at 1155 m asl. The topography of the regionvaries from undulating to mountainous relief. Bedrock comprisesgneiss and migmatite rocks belonging to the Guaxupé Complex.The climate in the region is a Cwb type (moderate humid

sub-tropical) according to the Köppen classification, with meantemperatures in the warmest and coldest months of 22� and15 �C. There are two well-defined seasons (warm and humidsummer; cold and dry winter), with a mean annual rainfall ofaround 1500 mm. The soils region moisture regime is udic and thetemperature regime is isothermic. The studied soil is located onthe higher areas (higher morphologies) where old erosionvestiges are still evident, possibly the South American or VelhasSurface (King, 1956), and it has a clay texture. The water pHincreases from 4.2 to 5.5 from the surface to the bottomof the profile (Table 1). The local vegetation is composed ofa sub-perennial tropical forest with occurrence of Cerrado patches(Silva and Vidal-Torrado, 1999), and the region is an ecotonebetween these two formations.

3. Material and methods

3.1. Soil description, physical properties and sampling

The profile was cut to a depth of 200 cm on the upper part ofa slope, and the soil exposed was described according to Santoset al. (2006). Loose samples were collected of each soil horizons.Samples were air-dried, and particle size analysis was carried outafter dispersion with 0.01 mol L�1 NaOH solution according toEmbrapa (1997). Soil bulk density was calculated using the amountof dry soil (grams) contained in 100 cm3 cylinders collected intriplicate from the field (Embrapa, 1997).

Table 2AMS radiocarbon assays.

Material dated Depth (cm) Conventional 14C Age BP 2-sigma 14C Cal age BP

Charcoala 20e40 210 � 30 180 � 36Charcoala 100e120 10,320 � 120 12,131 � 428

a UGAMS e The University of Georgia e Center for Applied Isotopes Studies.

Table 3Phytolith indices.

Sample/Depth (cm) Indices

Iph*100 IC*100 D/P

F1 (0e10) 0 16.1 0.2F2 (10e20) 27.8 21.4 0.1F3 (20e30) 56.7 20.8 0.1F4 (30e40) 18.8 27.7 0.01F5 (40e50) 47.1 34.0 0.01F6 (50e60) 46.9 31.5 0.1F7 (60e70 66.1 24.2 0.05F8 (70e80) 62.2 27.2 0.03F9 (80e90) 90.2 13.0 0.1F10 (90e100) 72 30.5 0.03F11 (100e110) 45.2 19.7 0.1

Fig. 2. Mircrophotography of Phtyoliths. I) Poaceae phytoliths morphotypes found in the umbric epipedon (in %). a,b) bilobates; c) bilobate - Chusqueatype (cf. Montti et al. 2009);d) rondel; e) bulliform; f,g) different types of elongate morphologies.; II) a, b) globular echinate (Arecaceae morphotypes); c, d) crater-shaped (Araucariaceae morphotypes);e, f) globular and g,h) irregularly shaped (both Dicotiledoneae morphotypes).III) Photomicrographs of other silica bodies and pollen found in the umbric epipedon. Pictures weretaken at 400� magnification.

M.R. Calegari et al. / Quaternary International 287 (2013) 47e55 49

Author's personal copy

3.2. Soil organic matter e SOM

The samples were taken in 10 cm increments to a depth of110 cm. The total organic carbon and d13C analyses were carried outat the Stable Isotopes Laboratory of the Centre for Nuclear Energy inAgriculture (CENA-University of Sao Paulo - Brazil). The Totalcarbon results are expressed in grams per kilogram of dry matter(g kg�1) for each horizon (Table 1) and the 13C results are expressedas d13C using the conventional d (&) notation, with an analyticalprecision of �0.2&:

d13C�&�

¼hRsample=Rstandard � 1

i1000

Where: Rsample refers to the 13C/12C ratio of the sample and Rstandardrefers to the 13C/12C ratio of the standard.

Soil organic matter AMS dating was performed at the AMSLaboratory of Centre for Applied Isotopes Studies, University ofGeorgia, United States, on charcoal fragments physically separatedfrom the soil by hand-picking and transformed in CO2 bycombustion at 14C Laboratory of CENA-USP (Brazil). Radiocarbonages are reported as 14C (1s) BP (present is AD 1950) normalized toa d13C of�25& VPDB and in calibrated years as cal (2s) BP using theprogram CALIB 5.0. Two sections of the epipedon, 20e40 cm and100e120 cm, were sampled, as the objective was the under-standing of umbric epipedon genesis (Table 2).

3.3. Phytolith sampling and extraction

The phytolith samples were taken in parallel to the SOM ones at10 cm increments to a depth of 110 cm. The presence of biologicalchannels (bioturbation) is common in Oxisol pedons, indicatingpossible mixture of materials and suggesting upward movementand translocation of materials along the profile. However, duringsampling the areas of the soil section with minimum bioturbationfeatures were selected to avoid mixed material that would result ininadequate samples.

Phytolithswere extracted from4 g of dry soil. The organicmatterwas oxidized with hydrogen peroxide at 30%, first at room temper-ature and then heated to 70 �C, followed by carbohydrate and ironoxides removal using sodium citrate-bicarbonate-dithioniteaccording to Mehra and Jackson (1960). The density separation ofphytoliths from the resulting fraction was performed using sodiumpolytungstate with a density of 2.3, according to Madella et al.(1998). The recovered fraction, which included phytoliths, diatomsand other silicate bodies, was mounted with immersion oil onmicroscope slides for 3D observations with a Zeiss Axioscopemicroscope at 400�magnification.

3.3.1. Phytolith classification and countingFor each sample at least 300 phytoliths with taxonomic signif-

icance were counted. A transect in each slide was counted todetermine the proportion of phytoliths and other particles in thefraction obtained after heavy liquid separation, according toCarnelli (2002). The identified phytoliths were named following theInternational Code for Phytolith Nomenclature (Madella et al.,2005) and then grouped in accordance to their taxonomicalsignificance (Mulholland, 1989; Twiss, 1992; Fredlund and Tieszen,1994; Alexandre et al., 1997; Alexandre et al., 1999; Runge, 1999;Parr and Watson, 2007):

(a) Poaceae (grasses)Pooideae or pooid (trapeziform and rondel)Panicoid or panicoid (bilobate and cross)Chloridoideae or chloridoid (sadde)

(b) Arecaceae/palms (gloubar echinate)(c) Dicotiledoneae/woody (globular, plate, papillae, irregular)(d) Araucariaceae (crater-shaped cell)

For the interpretation of palaeoenvironmental conditions, thefollowing phytolith indices were calculated (Table 3):

The Humidity-Aridity Index (Iph) (Diester-Haas et al., 1973;Alexandre et al., 1997) is based on the ratio of chloridoid versuschloridoid and panicoid phytoliths. High Iph values suggest openwoodlands and/or grasslands dominated by xerophytic Chlor-idoideae, indicating dry edaphic and/or climatic conditions.Conversely, low Iph values indicate the predominance of meso-phytic Panicoideae, suggesting more humid conditions (Alexandreet al., 1997).

The Climatic Index (Ic) (Twiss, 1987, 1992) is the ratio of pooidversus the sum of pooid, chloridoid and panicoid morphotypes.High values indicate the predominance of C3 Pooideae grasses,suggesting cold climatic conditions (Twiss, 1992).

The Tree Cover Density Index (D/P) (Alexandre et al., 1997, 1999;Barboni et al., 1999) consists of the D/P ratio, where D is the numberofdicotyledonphytoliths (globularmorphotypes) andP is thenumberof Poaceae phytoliths (pooids, chloridoids, panicoids, trichomes andbulliforms). High values indicate open vegetation, adapted to warmand dry climates, as in the African tropical and intertropical zones;lower values indicate forest vegetation with warm and wet climates(Alexandre et al., 1997, 1999; Barboni et al., 1999).

3.4. Statistical analysis

Two forms of statistical analysis were applied to over phytolithdata: Cluster Analysis (CA) and Principal Component Analysis

Table 4Phytolith morphotype frequencies from the hyper-developed umbric epipedon.

Sample/Depth (cm)

Angiospermae

Monocotiledoneae

Poaceae/Grasses

Panicoideae Pooideae Chloridoideae Cylindric Elongate Bulliform Trichome Clavate Hair

% % % % % % % % %

F1 (0-10 41 12 35 10 0 0 7 2 50 15 128 38 7 2 0 0 0 0F2 (10-20) 13 4 28 10 5 2 6 2 64 21 69 23 0 0 15 5 10 3F3 (20-30) 13 4 32 11 17 6 8 3 52 18 70 24 0 0 5 2 14 5F4 (30-40) 13 4 33 10 3 1 1 0.3 53 16 64 20 0 0 12 4 5 2F5 (40-50) 18 5 66 19 16 5 3 1 62 18 87 25 0 0 18 5 4 1F6 (50-60) 17 5 57 16 15 4 1 1 58 17 89 26 0 0 12 4 0 0F7 (60-70 20 6 53 16 39 12 1 0.3 66 20 106 32 0 0 18 5 0 0F8 (70-80) 14 4 58 18 23 7 1 0.3 60 19 112 35 0 0 24 7 5 2F9 (80-90) 4 1 29 10 37 12 0 0 32 11 137 45 0 0 7 2 16 5F10 (90-100) 7 2 71 21 18 5 0 0 47 14 135 41 0 0 26 8 2 1F11 (100-110) 17 5 39 12 14 5 0 0 63 20 118 38 1 0.3 0 0 9 3

a The number of taphonomised phytoliths is not considered when calculating the total sum of phytoliths.

M.R. Calegari et al. / Quaternary International 287 (2013) 47e5550

Author's personal copy

(PCA). For both analyses, Ward’s methodwas applied and for the CAthe Euclidian distance was used. The clustering was based onbinary grouping according to the degree of similarity between thesamples (Beebe and Kowalski, 1987). The dataset were expressed aspercentages.

The dataset was not normalized because the analyses were runon the original correlation matrix expressed in percentages. Ward’smethod and the squared Euclidean distance were used for CA.

4. Results

4.1. Soil morphology

The studied pedon is deep (for other profiles in this area it isbetween 700 cm and 1000 cm), and extremely well-drained. It hasa minimum 220 cm thick dark and organic carbon-rich A horizonincluding well expressed transitional AB and BA horizons, classifiedas humic by the SiBCS (Embrapa, 2006) and a very thick umbricepipedon by Soil Taxonomy (USDA, 1999). A recent 10 cm deepdeposit of reworked material (colluvium) was identified in theupper epipedon. However, as suggested by the soil position in thelandscape (upper third of the slope) and the morphology and Ti/Zrratio (Calegari, 2008), the colluvial process was not the primaryfactor causing the formation of this very thick epipedon. The Ti/Zrratio shows that the establishment and development (thickening)of the epipedons occurred after the older colluvial deposition.

Macroscopic charcoal fragments were identified in the pedon(between 20e40 and 100e120 cm) and dated at 180 � 36 cal BPand 12,131 � 428 cal BP (Table 1). These ages are consistent withthose found by Silva and Vidal-Torrado (1999) in another Machado(MG) soil located not far from the profile studied in this work. Thesoil charcoals represent a period of fires associated with drierclimatic spells during the Holocene (Pessenda et al., 1996, 2004).

The highest total carbon amount of 83.82 g kg�1 was found atthe top of the A2 horizon, between 20 and 30 cm depth (Table 2).Below 40 cm, total carbon values gradually decrease. Kaolinite (Kt)was themain claymineral in all soils where XRD analysis was done,followed by gibbsite (Gb) (Marques et al., 2011).

The bulk density in the umbric epipedon is lower than in the Bw(Table 1). This trend may be associated with the variations intexture and soil structure between horizons. The clay contentranged (top to bottom) from 498 g kg�1 to 650 g kg�1, and thesevalues are typical for this type of soil. Generally, the Humic Oxisolvaries from clayey to very clayey as a result of intense desilicifica-tion that precedes organic carbon accumulation (Marques et al.,2011). In this epipedon there are clear vertical channels (pedotu-bules). They are filled with small to very small granular aggregates,well-rounded and reddish/dark materials originating from

overlying horizons. This type of bioturbation has been consideredto be the result of intense activity of termites (Marques et al., 2011).These features are particularly well-preserved in Brazilian Ferral-sols with high clay content (Silva and Vidal-Torrado, 1999; Marqueset al., 2011).

4.2. Phytolith assemblages and d13C values

The main phytolith types that were identified and counted arepresented in Fig. 2 and Table 4. The amount of phytoliths per gramof soil is variable, with two important peaks at 30e40 cm and at100e110 cm (Table 4).

Hierarchical cluster and principal components (PCA) analyseswere performed on the phytolith assemblages expressed inpercentages. The dendrogram shows that the samples cluster inthree groups (Fig. 3) that are in agreement with the pedologicalhorizons previously defined by morphological conformity (colour,texture, structure, porosity, etc.). Group I is composed by thesamples from the horizons AB, A3 and the base of A2, group II iscomposed by the upper part of the A2 horizon, and group IIIcorresponds to the modern samples from the top of the soil.

The PCA first axis accounts for a 33.2% of the total variance and itis defined by bulliform, Dicotiledoneae, Chloridoideae, Panicoideaeand Pooideae phytoliths for the positive values and palm andAraucaria phytoliths for the negative values (Fig. 4). This axis can beconsidered a proxy for the temperature, because the Araucariaforest ecological limit in central Brazil is within a mean wintertemperature of <10� (Ledru, 1993).

The PCA second axis expresses 27% of the total variance, and it isdefined by Dicotiledoneae and Panicoideae for the positive valuesand Chloridoideae and Pooideae for the negative values. This axiscan be considered a proxy for humidity, varying between mesic toxeric conditions. Woody/Dicotiledoneae plants as well as Pan-icoideae grasses need higher humidity while Choridoideae aremore adapted to dry conditions.

On the basis of the similarity expressed by the statistical anal-ysis, three phytolith zones were identified (Fig. 5), from bottom totop:

Zone I (from ca. 12,131 � 428 cal BP to ca. 6000 cal BP), sub-divided as:

Sub-zone Ia: This sub-zone includes horizon AB andsub-horizon A3 (110e70 cm). The soil has a dark reddish browncolor (5 YR e 2.5/2), clay texture (Table 1) and a large amount ofvertical pedotubules that extend through the transition betweenAB and BA horizons. Charcoal fragments were found in largequantities along this transition. The phytolith assemblages aredominated by Poaceae morphotypes that represent, as an average,88% of the identified assemblage (Table 2). Dicotiledoneae (globular

Angiospermae Gymnospermae Not identified Taphonomiseda Amount ofPhytolith/g.soil

Monocotiledoneae Dicotiledoneae (trees and shrubs only) AraucariaceaeAraucaria

Arecaceae PalmsGlobularEchinate

Globularpsilate/rugose

Papilla Irregular Crater-shaped

% % % % % % %

18 5 42 12 2 1 0 0 0 0 11 3 33 10 12,800,00071 24 8 3 2 1 3 1 5 2 3 1 23 8 2,997,06952 18 8 3 2 1 6 2 6 2 9 3 52 18 69,226,667

135 41 1 0 4 1 0 0 1 0.3 2 1 17 5 122,716,00035 15 2 1 5 1 7 2 3 1 6 2 6 2 95,480,00058 17 16 5 0 0 11 3 1 0.3 5 1 6 2 36,055,55611 3 10 3 8 2 0 0 0 0 0 0 10 3 27,463,33311 3 7 2 0 0 1 0.3 0 0 6 2 16 5 35,200,00024 8 16 5 1 0 0 0 0 0 0 0 20 7 38,390,0007 2 6 2 2 1 0 0 0 0 7 2 14 4 77,000,000

21 7 29 9 0 0 6 0 0 0 0 0 37 12 316,800,000

M.R. Calegari et al. / Quaternary International 287 (2013) 47e55 51

Author's personal copy

morphotypes) frequency ranges between 2% and 9%, and palms(Arecaceae e globular echinate morphotypes) have values between2% and 8%. Iph is more than 45%; Ic is less than 30% (with the lowestvalue of 19% at the bottom); D/P varies between 0.19 and 0.03(Table 3). Between 120 cm and 70 cm d13C values indicate a smallbut progressive isotopic depletion from �23.2& to �24.2&, sug-gesting a mixture of C3 and C4 plants but with predominance of C3(most likely woody elements) (Pessenda et al., 1996, 1998; Gouveiaet al., 1999, 2002).

Sub-zone Ib corresponds to the top of sub-horizon A3 and thebottom of A2 (70-40 cm). Darker colour and a large amount ofcharcoal fragments characterize this zone. Biological activity is alsointense and pedotubules and galleries filled with material of lightercolour (probably from the Bw horizon) are common. The sub-zoneis separated on the basis of the frequency of globular echinatephytoliths (15%e415%) produced by palms (Arecaceae) as well asthe appearance of crater-shaped phytoliths (1%e2%), produced byAraucaria trees. However, the phytolith assemblage is still domi-nated (76%) by Poaceae phytoliths (of which 5% are Panicoideae,18% are Pooideae and 4% are Chloridoideae). Dicotiledoneae typesaccount for 6% of the assemblage. The charcoal fragments found inthis sub-zone are at similar depth as those found by Silva and Vidal-Torrado (1999) in a nearby soil with similar characteristics and with

an age of 6103 � 113 cal BP The index Iph is 47%e48%, Ic variesbetween 28% and 34%, and D/P varies between 0.01 and 0.1(Table 3). A d13C value of �23.4&, presents the same tendency ofsub-zone Ia, and suggests a mixture of C3 and C4 plants, withpredominance of C3 (Fig. 6).

Zone II corresponds to the A2 sub-horizon (40-10 cm) Colourand texture are similar to those of sub-horizon A3, although thestructure is slightly more granular. Phytoliths assemblage showeda very noticeable decrease in Poaceae morphotypes (the sum ofPooideae, Cloridoideae and Panicoideae becomes only 17% of theassemblage) (Table 2). In this zone the Araucariaceae types aremore common, reaching a frequency of 2% but the Palms (Areca-ceae) types show a more considerable increase, reaching 41% of thephytolith sum at the bottomof this zone. The values of the phytolithindices Iph are generally less than 20% with the exception of the40-30 cm sample (56.9%); Ic less than 30% and D/P less than 0.1(Table 3). The d13C value (�23.0&) is a mixture of C3 and C4 plants,with predominance of C3 and the more depleted values between40 cm and 20 cm (�25.0&) suggest a progressive predominance ofarboreal vegetation. Several diatom exoskeletons (Fig. 2-III) andPteridophytae spores were observed in the microscope slides, allindicating more humid environmental conditions. Charcoal frag-ments lines as well as sparse fragments were observed in the upper

Fig. 3. Cluster analysis (Ward method) for the phytoliths assemblage of the umbric epipedon.

Fig. 4. Principal component analysis of the phytoliths assemblage from the umbric epipedon. (A) Ellipses and points dispersion diagram for each phytolith zone. (B) PrincipalComponents Regression.

M.R. Calegari et al. / Quaternary International 287 (2013) 47e5552

Author's personal copy

part of this zone (between 30 and 20 cm), and the charcoal wasdated 180 � 36 cal BP, probably reflecting anthropogenic activitiesduring the colonial period (Silva and Vidal-Torrado, 1999).

Zone III corresponds to the deposit of the current vegetation (anopen forest) and shows d13C value of �23.0&. The charcoal frag-ments found in this zone were dated to the modern age, after 1950.

In this zone an increase of Poaceae phytoliths in comparison withZone II was observed. The phytolith indices are: Iph 0, Ic 16.1% andD/P 0.2 (Table 3 and Fig. 6).

5. Discussion

The pedogenetic history of Brazilian Oxisols seems to haverecorded some climatic variations as well as erosion cycles sincethe Neogene (Müggler, 1998; Schaefer, 2001). However, the resultssuggest low climatic variability since 12,000 cal BP, as evident bythe isotopic signal of the organic matter accumulated in the humichorizon, derived from plant formations composed by a mixture ofC3 and C4 plants. The interactions between different forms of

Fig. 6. Range of d13C values in the umbric epipedon.

Fig. 5. Phytolith spectrum and indexes from the umbric epipedon.

Fig. 7. Details of the biological channels in transitions horizons AB/BA of HumicHapludox profile (Machado, Minas Gerais State, Brazil).

M.R. Calegari et al. / Quaternary International 287 (2013) 47e55 53

Author's personal copy

carbon and poorly crystalline aluminum represent a mechanism ofcarbon protection in the humic epipedon that results in higher OMaccumulation (Marques et al., 2011).

The signal in Zone I (between 12, 131 � 428 cal BP and6103 � 113 cal BP), the base of the humic horizon (umbric epi-pedon), is of an open savannah with arboreal elements and a grasscover mainly composed of C3 taxa. This vegetation was associatedwith a drier-than-present climate, also identified in the region byLedru (1993), Ledru et al. (1994), Alexandre et al. (1999) andPessenda et al. (1996, 2004). The first indication of a more signifi-cant abundance of arboreal elements in the vegetation cover and ofan increase in humidity is observed from around 6000 cal BP(50e60 cm depth e top of Zone Ib) with the appearance of Arau-caria phytoliths (crater-shaped).

The low values of the D/P index in Zone I are similar to the onesfrom shrub steppes observed in Africa by Barboni et al. (1999).Alexandre et al. (1999) interpret theD/P values between 1 and 1.3 inan Oxisol from Salitre (Minas Gerais, Brazil), more or less contem-porary to Zone I of the current profile, as open vegetationwith sometrees and shrubs. These data are also in agreement with the resultspresented by Silva and Vidal-Torrado (1999). This zone hasa common presence of biological channels (bioturbation) between120 and 90 cm (Fig. 7). The occurrence of bioturbation in BrazilianOxisols, particularly thosewith high clay content, is rather commonbecause these soils can preserve these features for long periods(Boulet et al., 1995; Behling et al., 1998; Gouveia and Pessenda,2000; Schaefer, 2001). The linear increase with depth of thecorroded phytoliths together with the decrease of d13C and organiccarbon content are consequence of prolonged pedogeneticprocesses. These characteristics are common in many tropical andsub-tropical soils (Parton et al., 1987) and are observed in themajority of the isotopic datasets fromBrazilian soils (Vidotto, 2008).

In Zone II, the phytolith assemblages show an increase of thecrater-shaped (Araucariaceae) and of globular echinate (Arecaceaee palms) morphotypes together with a decrease of grass (Poaceae)phytoliths as well as an isotopic depletion (Fig. 6). The signal fromthis trend is reinforced by the presence of Dicksonia sellowianaHook. spores and diatom exoskeletons. This evidence points to anincrease in humidity from around 6000 cal BP that was alreadyobserved at the top of Zone I. The current results are in agreementwith earlier studies that highlighted a humid phase in Central Brazilduring the Mid-Late Holocene with the establishment of climaticconditions similar to the present and the formation of forestvegetation (Oliveira, 1992; Ledru, 1993; Pessenda et al., 1996, 1998,2004; Scheel-Ybert et al., 2003).

Finally, the phytoliths from Zone III represent the modern dayvegetation, an ecotone of Sub-perennial Tropical Forest and Cerrado(BRASIL, 1983; Silva and Vidal-Torrado,1999) with species typical ofCerrado such as Casearia sylvestris Sw. and Piptadenia gonoacantha(Mart.) Macbr. The presence of charcoal fragments (180� 36 cal BP)and charred phytoliths indicate fire events during the 18th century.Silva and Vidal-Torrado (1999) interpreted these last fire episodesas resulting from the initial occupation of the area during thecolonial period.

6. Conclusions

This is the first time that an Oxisol with a humic horizon hasbeen studied in detail using two powerful and complementarypalaeoenvironmental techniques. The association of d13C valuesand phytolith records has shown that it is extremely important torely on a combined set of evidence to refine understanding ofenvironmental and climatic evolution of tropical and sub-tropicalareas. The poor pollen preservation in this kind of terrestrialrecord makes it even more fundamental to move to an

interdisciplinary approach and to take advantage of all the availablerecords preserved in these deposits.

On the basis of d13C and phytolith records, it is likely that theorganic matter accumulated in the humic horizon of this profilederived from plant formations composed by a mixture of C3 and C4plants, with a certain increase in C3 plant contribution from theMiddle Holocene. The index values characterizing the studied soilare similar to those of soils from African tropical regions (Alexandreet al., 1997; Barboni et al., 1999, 2007) and to other studies in Brazil(Alexandre et al., 1999). They indicate dry to humid mesothermicclimatic conditions suitable for the development of a mixture ofvegetation with C3 and C4 plants, but with a predominance of C3plants. Despite the general agreement with the values from othergeographical areas, it is necessary to improve the accuracy of theindices for the tropical and sub-tropical vegetation of SouthAmerica.

Finally, there is also the need for additional studies of phytolithassemblages from tropical modern plants in South America toobtain a more precise taxonomic understanding of the phytolithassemblages that will allow distinguishing vegetation variability inrelation to topography, humidity and temperature.

Acknowledgements

The authors express their gratitude toProf. F.MaciasVásquez andProf. J. L. Ottero of the Universidad de Santiago de Compostela(Spain) for kindly making available the laboratories for phytolithchemical treatment and extraction. Thepresentworkwaspart of thedoctoral research of the first author and she wish to thanks FAPESPfor financial support, CAPES-MECD/DGU for providing the overseasscholarship and the personnel of the Palaeoecology and PlantPalaeoeconomy Laboratory of the Institució Milà i Fontanals of theSpanish National Research Council (CSIC) in Barcelona for the warmwelcome and technical support during her training internship.

References

Absy, M.L., Cleef, A.M., Fournier, M., Martin, L., Servant, M., Sifeddine, A., Ferreira daSilva, M., Soubies, F., Suguio, K., Turcq, B., Van Der Hammen, T., 1991. Mise enévidence de quatre phases d’ouverture de la forêt dense dans le sud-est del’Amazonie au cours des 60,000 derniéres anneés. Premiérecomparaison avecd’autres régions tropicales. Comptes Rendus de l’Academie des Sciences deParis 312 (II), 673e678.

Alexandre, A., Meunier, J.-D., Lczine, A.-M., Vincens, A., Schwartz, D.A., 1997.Phytoliths: indicators of grassland dynamics during the late Holocene inintertropical Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 136,213e229.

Alexandre, A., Meunier, J.-D., Mariotti, A., Soubies, F., 1999. Late Holocene phytolithand carbon-isotope record from a Latosol at Salitre, south-central Brazil.Quaternary Research 51, 187e194.

Argant, J., 1990. Climat et environnement au Quaternaire dans le Bassin du Rhôned’après les données palynologiques. Document Laboratoire Géologique de Lyon,Lyon. n.111, p.199.

Barboni, D., Bonnefille, R., Alexandre, A., Meunier, J.D., 1999. Phytoliths aspaleoenvironmental indicators, West Side middle Awash Valley, Ethiopia.Palaeogeography, Palaeoclimatology, Palaeoecology 152, 87e100.

Barboni, D., Bremond, L., Bonnefille, R., 2007. Comparative study of modernphytolith assemblages from inter-tropical Africa. Palaeogeography, Palae-oclimatology, Palaeoecology 246, 454e470.

Beebe, K.R., Kowalski, B.R., 1987. An introduction to multi-variate calibration andanalysis. Analytical Chemistry 59, 1007Ae1017A.

Behling, H., Lichte, M., Miklós, A.W., 1998. Evidence of a forest free landscape underdry and cold climatic conditions during the last glacial maximum in theBotucatu region (São Paulo State), Southeastern Brazil. In: Rabassa, J.,Salemme, M. (Eds.), Quaternary of South America an Antarctic Peninsula, vol. 11.A.A. Balkema, Rotterdam, pp. p99e110.

Borba-Roschel, M., Alexandre, A., Varajao, A.F.D.C., Meunier, J.D., Varajao, C.A.C.,Colin, F., 2006. Phytoliths as indicators of pedogenesis and paleoenvironmentalchanges in Brazilian cerrado. Journal of Geochemical Exploration 88 (1/3),172e176. Amsterdam.

Boulet, R., Pessenda, L.C.R., Telles, E.C.C., Melfi, A., 1995. Une evaluation de la vitessede l’accumulation de matière par la faune du sol à partir de la datation dêscharbones et d’humine du sol. Exemple dês versants du lac Campestre, Salitre,

M.R. Calegari et al. / Quaternary International 287 (2013) 47e5554

Author's personal copy

Minas Gerais, Brasil. In: Comptes Rendus de L’Academie dês Sciences, vol. 320,pp. 287e294. Série IIA, Paris.

BRASIL. Ministério das Minas e Energia. Secretaria Geral, 1983. Projeto RADAM-BRASIL. Folhas SF23/24, Rio de Janeiro/Vitória: Geologia, geomorfologia,pedologia, vegetação e uso Potência da Terra. Rio de Janeiro: Ministério dasMinas e Energia, 1983. 780p. (Levantamento dos Recursos Naturais, 32).

Bremond, L., Alexandre, A., Hély, C., Guiot, J., 2005. A phytolith index as a proxy oftree cover density in tropical areas: calibration with Leaf Area Index alonga forestesavanna transect in southern Cameroon. Global and Planetary Change45 (4), 277e293.

Calegari, M.R., 2008. Ocorrência e Significado paleoambiental do horizonte Ahúmico em Latossolos. Escola Superior de Agricultura Luiz de Queiroz,Piracicaba. 259p (Tese de Doutorado).

Carnelli, A., 2002. Long term dynamics of the vegetation at the subalpine-alpineecocline during the Holocene: comparative study in the Aletsch region, Vald’Arpette, and Furka Pass (Valai, Switzerland). Terre & Environnement 40, 349.

Diester-Haas, L., Schrader, H.J., Thiede, J., 1973. Sedimentological andpaleoclimatological investigations of two pelagic ooze cores off Cape Barbas,North-West Africa. Meteor Forsch-Ergebnisse, C 16, 19e66.

Embrapa, 1997. Empresa Brasileira de Pesquisa Agropecuária e Embrapa. In:Manual de métodos de análises de solo, vol. 2. Centro Nacional de Pesquisa deSolos, Rio de Janeiro, p. 212.

Embrapa, 2006. Empresa Brasileira de Pesquisa Agropecuária e Embrapa. CentroNacional e Pesquisa em Solos. Sistema Brasileiro de Classificação de Solos,Brasilia: Embrapa-SP; Rio de Janeiro. 306p.

FAO-UNESCO, 1981. Soil Map of the World (1:5.000.000) v.1e10. UNESCO, Paris.Fredlund, G.G., Tieszen, L.L., 1994. Modern phytolith assemblages from the North

American great Plains. Journal of Biogeography 21, 321e335.Gouveia, S.E.M., Pessenda, L.C.R., 2000. Datation par Le 14C de Charbons inclus dans

lê sol pour l’etude du role de la remontée biologique de matiére et ducolluvionnement dans la formation de latosols de l’etat de São Pauloi, Brésil. In:Comptes Rendus de L’Académie dês Sciences, vol. 330, pp. 133e138. Série IIA,Paris, n 2.

Gouveia, S.E.M., Pessenda, L.C.R., Aravena, R., Boulet, R., Roveratti, R., Gomes, B.M.,1997. Dinâmica de vegetações durante o Quaternário recente no sul do Ama-zonas indicada pelos isótopos do carbono (12C, 13C e 14C). Geochimica Brasi-liensis 11 (3), 355e367.

Gouveia, S.E.M., Pessenda, L.C.R., Boulet, R., Aravena, R., Scheel-Ybert, R., 1999.Isotopos do carbono dos carvões e da matéria orgânica do solo em estudos demudanças de vegetação e clima no Quaternário e da taxa de formação de solosdo estado de São Paulo. Anais da Academia Brasileira de Ciências 71 (4),969e980.

Gouveia, S.E.M., Pessenda, L.C.R., Aravena, R., Boulet, R., Scheel-Ybert, R.,Bendassoli, J.A., Ribeiro, A.S., Freitas, H.A., 2002. Carbon isótopos in charcoal andsoils in studies of paleovegetation na climate changes during the latePleistocene and the Holecene in the southeast and centerwest regions of Brazil.Global and Planetary Change 33 (1e2), 95e106. Amsterdam.

Kampf, N., Klamt, E., 1978. Mineralogia e Gênese de Latossolos (Oxisols) e SolosPodzólicos da Região Nordeste do Planalto Sul-Riograndense. Revista Brasileirade Ciência do Solo, Viçosa 2, 68e73.

King, L.C., 1956. A geomorfologia do Brasil Oriental. Revista Brasileira de Geografia18 (2), 147e265. Rio de Janeiro.

Kondo, R., Iwasa, Y., 1981. Opal phytoliths in humic yellow laterosols and yellowlaterosols in the Amazon region. Research Bulletin Obihiro University 12,231e239.

Ledru, M.-P., 1993. Late quaternary environmental and climatic changes in CentralBrazil. Quaternary Research 39, 90e98.

Ledru, M.P., Behling, H., Fournier, M., Martin, L., Servant, M., 1994. Localisation de laforêt d’Araucaria du Brésil au cours de l’holocène e implications paléoclima-tiques. Comptes Rendus de l’Académie des Sciences, Sciences de la vie 317 (III),517e521.

Lepsch, I.F., Buol, S.W., 1986. Oxisol-landscape relationship in Brazil. In:International Soil Classification Work Shop, 13, Campinas. 1986. Part1: papers.Campinas: SNLCS/EMBRAPA. University of Puerto Rico, USA, pp. 174e189.

Madella, M., Powers-Jones, A.H., Jones, M.K., 1998. A Simple method of Extraction ofopal phytoliths from Sediments using a Non-Toxic heavy liquid. Journal ofArchaeological Science 25, 801e803.

Madella, M., Alexandre, A., Ball, T., 2005. International Code for phytolithNomenclature 1.0. Annales of Botany 96 (2), 253e260.

Marques, F.A., Calegari, M.R., Vidal Torrado, V., Buurman, P., 2011. Relationshipbetween soil oxidizable carbon and physical, chemical and mineralogicalproperties of Umbric Ferralsons. Revista Brasileira de Ciência do Solo 35, 25e40.

Martinelli, L.A., Pessenda, L.C.R., Espinoza, E., Camargo, P.B., Telles, E.C., Cerri, C.C.,Victoria, R.L., Aravena, R., Richey, J., Trumbore, S., 1996. Carbon-13 variationwith depth in soils of Brazil and climatic change during Quaternary. Oecologia106, 376e381.

Mehra, O.P., Jackson, M.L., 1960. Iron oxide removal from soils and clays bya dithioniteecitrate system buffered with sodium bicarbonate. Clays and ClayMinerals 7, 317e327.

Montti, L., Honaine, M.F., Osterrieth, M., Ribeiro, D.G., 2009. Phytolith analysis ofChusquea ramosissima Lindm. (Poaceae: Bambusoideae) and associated soils.Quaternary International 193 (1-2), 80e89.

Müggler, C.C. 1998. Polygenetic oxisols on Tertiary surfaces, Minas Gerais, Brazil-soilgenesis and landscape development. Thesis (PhD), The Netherlands: Wage-ningen Agricultural University, Wageningen, 185p.

Mulholland, S.C., 1989. Phytolith shape frequencies in North Dakota grasses:a comparison to general patterns. Journal of Archaeological Science 16, 489e511.

Nakashima, P., 1973. Estudo das formações superficiais na área de Campinas e

Viracopos (SP): suas relações com a evolução geomorfológica e os solos. 149p.Dissertação, Mestrado na área de Geografia Física, Faculdade de Filosofia, Letrase Ciências, Universidade de São Paulo, São Paulo.

Oliveira, P.E., 1992. A Palynological record of late quaternary vegetation and climaticchange in southeastern Brazil, PhD Dissertation, Ohio State University,Columbus, 238p.

Parr, J., Watson, L., 2007. Morphological characteristics observed in the leafphytoliths of selected Gymnosperms of eastern Australia. In: Madella, M.,Zurro, D. (Eds.), Plants, People and Places- Recent Studies in Phytoliths Analysis.Oxbow Books, p. 272.

Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1987. Analysis of factors controllingsoil organic matter levels in Great Plains grasslands. Soil Science Society ofAmerica Journal 51, 1173e1179.

Pessenda, L.C.R., Aravena, R., Melfi, A.J., Telles, E.C.C., Boulet, R., Valencia, E.P.E.,Tomazello, M., 1996. The use of carbon isotopes (C-13, C-14) in soil to evaluatevegetation changes during the Holocene in Central Brazil. Radiocarbon 38 (2),191e201.

Pessenda, L.C.R., Gomes, B.M., Aravena, R., Ribeiro, A.S., Boulet, R., Gouveia, S.E.M.,1998. The carbon isotope record in soils along a forestecerrado ecosystemtransect: implications for vegetation changes in the Rondonia state, south-western Brazilian Amazon region. The Holocene 8 (5), 631e635.

Pessenda, L.C.R., Gouveia, S.E.M., Aravena, R., Boulet, R., Valencia, E.P.E., 2004.Holocene fire and vegetation changes in southeastern Brazil as deduced fromfossil charcoal and soil carbon isotopes. Quaternary International 114, 35e43.

Pessenda, L.C.R., Gouveia, S.E.M., Freitas, H.A., Ribeiro, A.S., Aravena, R.,Bendassoli, J.A., Ledru, M.P., Scheel-Ybert, R., 2005. Isótopos do carbono e suasaplicações em estudos paleoambientais. In: De Oliveira, P., Suguio, K.,Gouveia, C. (Eds.), O Quaternário No Brasil. Holos, São Paulo, pp. 75e93.

Piperno, D.R., 2006. Phytoliths: a Comprehensive Guide for Archaeologists andPaleoecologists. AltaMira Press, Lanham MD.

Queiroz Neto, J.P., Castro, S.S., 1974. Formações Superficiais e Latossolos Vermelho-Amarelo Húmico na Área de Bragança Paulista. Estado de São Paulo, Brasil. In:Congresso Brasileiro de Geologia, vol. 18, pp. 65e83. Porto Alegre. Anais. PortoAlegre: SBG.

Runge, F., 1999. The opal phytolith inventory of soil in central Africa e quantities,shapes, classification, and spectra. Review of Palaeobotany and Palynology 107,23e53.

Santos, R.D., Lemos, R.C., Santos, H.G., Ker, J.C., Anjos, L.H., 2006. In: Viçosa, M.G.(Ed.), Manual de descrição e coleta de solo no campo, vol. 5. Sociedade Brasi-leira de Ciência do Solo, p. 100.

Schaefer, C.E.R., 2001. Brazilian latosols and their B horizon microstructure aslong-term biotic constructs. Australian Journal Soil Research 39, 909e926.Melbourne.

Scheel-Ybert, R., Gouveia, S.E.M., Pessenda, L.C.R., Aravena, R., Coutinho, L.M.,Boulet, R., 2003. Holocene palaeoenvironmental evolution in the São PauloState (Brazil) based on anthracology and soil d13C analysis. The Holocene 13 (1),73e81.

Silva, A.C., Vidal-Torrado, P., 1999. Gênese dos Latossolos Húmicos e sua relação coma evolução da paisagem numa área cratônica do Sul do Minas Gerais. RevistaBrasileira de Ciência do Solo 23, 329e341.

Twiss, P.C., 1987. Grass opal phytoliths as climatic indicators of the Great PlainsPleistocene. In: Johnson, W.C. (Ed.), Quaternary Environments of Kansas. KansasGeological Survey Guidebook, vol. 5, pp. 179e188.

Twiss, P.C., 1992. Predicted world distribution of C3 and C4 grass phytolihs. In:Rapp Jr., G., Mulholland, S.C. (Eds.), Phytolith Systematics. Emerging Issues,Advances in Archaeological and Museum Science, vol. 1. Plenum Press,pp. 113e128.

USDA, 1999. Soil Taxonomy. A Basic System of Soil Classification for Making andInterpreting Soil Surveys. In: Soil Survey Staff, Handbook, second ed., vol. 436.United States Government Printing Office, Washington, DC.

Vidotto, E., 2008. Reconstrução paleoambiental (vegetação e clima) no ParqueEstadual da Ilha do Cardoso e SP durante o Quaternário tardio. Doutorado emEnergia Nuclear na Agricultura e no Ambiente. Centro de Energia naAgricultura, Universidade de São Paulo, Piracicaba. 210p. Tese.

M.R. Calegari et al. / Quaternary International 287 (2013) 47e55 55